Medical Imaging Methods, in Brief - Universitetet i oslo€¦ · ter responsible for central vision] Slide 26: Digital subtraction angiography • Angiographyimages: made while injecting

INF-GEO4310: Lecture Notes on Medical Imaging, Autumn 2012

Medical Imaging Methods, in Brief

Sven Peter NäsholmDepartment of Informatics, University of Oslo

Autumn semester, [email protected]

Office phone number: 22840068

Slide 2: Lecture overview

1. Medical imaging coordinate naming

2. X-ray medical imaging

• Projected X-ray imaging

• Computed tomography (CT) with X-rays

3. Nuclear medical imaging

4. Magnetic resonance imaging (MRI)

5. (Ultrasound imaging covered in previous lecture)

Slide 3: Medical imaging coordinates

The anatomical terms of location

• Superior / inferior, left / right, anterior / posterior:

Note: left / right is seen from the view of the patient!


Slide 4: Medical imaging planes

• Axial plane = transverse plane: perpendicular to the bodylong axis

• Sagittal: bisects the left from the right side [from new latinsagitta = arrow]

• Coronal: bisects the front from the back [from latin corona= crown]

Slide 5: Medical imaging planes

Coordinate system positive direction conventions

• 3 letters are used to indicate sequence and orientation ofthe (x, y, z) axes, e.g. “LSA” coordinates used in computedtomography (CT). Abbreviation letter codes:

superior (S)

inferior (I)

right (R)

anterior (A)

posterior (P)

left (L)S superior I inferiorL left R rightA anterior P posterior

• “LSA” thus indicates:

– x axis goes from left (L) to right

– y axis goes from superior (S) to inferior

– z axis goes from anterior (A) to posterior

• Other common coordinate system: “RPI”.

• What has to be done before combining images in e.g. LSAcoordinates with images in RPI coordinates?

Slide 6: Invasive / non-invasive

• Invasive imaging techniques:

– Optics inside the body, e.g. endoscope [greek: endo =inside]


– Open surgery

– Cameras in minimally invasive surgery, e.g. “camerapills” in veins or digestive tract

• Non-invasive imaging (Medical imaging) considerations:

– Planar projections vs. cross-sectional images

– Static images vs. dynamic series of images (“film”)

Slide 7: Classes of physical signals/processes

Non-invasive medical imaging: 4 different kinds of physical pro-cesses that generate signals:

• Ultrasound back-scattering / reflection

• X-ray transmission

• γ ray emission from radioisotopes

• Spin precession in magnetic fields

Slide 8: Imaging modalities

Method information carrier note

Ultrasound imaging high-frequency pressure waves backscattering / reflection imagingX-ray imaging (projection radiography)

ionizing radiationtransmission imaging

Computed tomography: CT, C-armNuclear imaging: SPECT, PET emission imagingMagnetic resonance imaging (MRI) nuclear spin precession magnetic resonance imaging

Slide 10: X-ray medical imaging

• Radiography: use of ionizing electromagnetic radiation toview objects

• Radiotherapy: use of ionizing electromagnetic radiation tocure diseases (thus not an imaging modality)

Slide 11: X-ray medical imaging

Brief historical retrospect

• Radiography started in 1895: discovery of X-rays by Wilhelm Conrad Röntgen (1845-1923,died from instetine cancer)

• First published picture using X-rays is of AnnaBerthe Röntgen’s hand in the paper “On ANew Kind Of Rays” (Über eine neue Art vonStrahlen)

• Awarded the first Nobel Prize in Physics(1901) In recognition of the extraordinary ser-vices he has rendered by the discovery of theremarkable rays subsequently named after him

• X-rays were put to diagnostic use very early,before the dangers of ionizing radiation werediscovered


Slide 12: Energetic electrons creating X-rays

Characteristic radiation

• Energetic electron collides with and ejects K-shell electron

• K-shell “hole” is filled by electron from the L,M, or N shells

• Produces characteristic spectral lines corre-sponding to the energy differences between theshells

Bremsstrahlung (braking / deceleration radiation)

• Energetic electron interacts with nucleus ofatom

• Deceleration causes loss of energy

• Continuous spectrum. Peaking @ anode-to-cathode potential

Slide 13: Conventional X-ray source

• Electrostatic lens to focus elec-tron beam onto a small spot @anode

• Anode designed to dissipateheat from focused electrons:

– Mechanically spun to in-crease the area heated bythe beam.

– Cooled by circulatingcoolant.

• Anode angled to allow escapeof some of the X-ray photonswhich are emitted essentiallyperpendicular to the directionof the electron current

• Anode usually made of tung-sten (W) or molybdenum (Mo)

• Window designed for escape ofgenerated X-ray photons


Slide 14: Image contrast

• Image contrast: caused byvarying absorption in object

• Absorption depends on atomicnumber Z:

– Metals distinguished fromtissues

– Bones distinguished byhigh Ca-content

• Absorption grows with Z2 dueto binding energy of inner elec-trons growing with Z2

• Absorption thus depends onprojected mass density:

– Lungs and air passagesform good contrast im-ages by density difference

– Higher water content(e.g., pneumonia) easilydetected

Slide 15: Projected X-ray transmission images

• Expose object to X-rays, capture“shadow”

• Produces a 2-D projection of a 3-D ob-ject

• “Shadow” may be converted to lightusing a fluorescent screen

• Image is then captured on either

– photographic film

– phosphorus screen to be “read”by laser

– matrix of detectors (digital radio-graphy)

• Projection radiography uses X-rays in different amounts andstrengths depending on what body part is being imaged:

– Hard tissues (bone) require a relatively high energy pho-ton source

– Soft tissues seen with same machine as for hard tissues,but with “softer” or less-penetrating X-ray beam

Slide 16: X-ray detectors

Ordinary photographic film


• Very inefficient, only 1–2% of radiation stopped

• Requires unnecessarily large X-ray dose to patient

Intensifying screens on both sides of film

• Phosphorus transforms x-ray photons into light photons

• Two types of luminescence

– Fluorescence: emission within 10−8 s of excitation

– Phosphorescence: emission delayed and extended

• Conversion efficiency: 5–20%

X-ray image intensifiers (XRIIs)

• Input window of aluminum or tita-nium

• X-ray photons absorbed by phospho-rus

• Channeled toward photocathode

Slide 17: Mammography

• X-ray examination used in diagnostic screening for breastcancer

• Used by radiologist or surgeon before biopsy (removal of cellsor tissue for examination), or lumpectomy (surgery to removee.g. a tumor)

• Radiation used for mammography tends to have a lower pho-ton energy than that used for bone and harder tissues

• Globally, breast cancer is the most lethal form of cancer forwomen (106 cases/year)

• US and Canada has highest incidence rate (100 per 100.000),but low mortality rate (19 per 100.000)

• Norway: incidence rate 75 per 100.000, mortality 16.1

Slide 18: True / False positives / negatives

True positive (TP)

Patient with cancer and positive test

True negative (TN)

Healthy patient and get negative test

False positive (FP)Healthy patient, but positive test

False negative (FN)Patient with cancer, but negative test

Good to have: TP & TN Bad to have: FP & FN


Slide 19: Sensitivity and specificity

Sensitivity

The dataset portion that tested positive out of all the positivepatients tested: Sensitivity = TP/(TP+FN)

• The probability that the test is positive given that the patientis sick

• The higher the sensitivity, the fewer decease cases go unde-tected

Specificity

The dataset portion that tested negative out of all the negativepatients tested: Specificity = TN/(TN+FP)

• The probability that a test is negative given that the patientis not sick

• Higher specificity means that a smaller percentage of healthypatients are labeled as sick

Slide 20: ROC curves

ROC = Receiver operating characteristic

• Plot of sensitivity vs. (1−specificity) fora binary classifier, as the discriminationthreshold is variedSame as: true positive rate vs falsepositive rate [TP/(TP+FN) vs.FP/(TN+FP)]

• Best possible method would give a pointin upper left corner of plane

– 100% sensitivity and specificity rep-resents the perfect classification.

• Result equivalent to random guessingwould lie on the diagonal.

Slide 21: Positive and negative predictive values

Positive predictive value (PPV)

The probability that a patient is sick, giventhat the result of the test was positive:PPV = TP/(TP+FP)

Negative predictive value (NPV)

The probability that a patient is not sick,given that the test result was negative:NPV = TN/(TN+FN)

• Assume a classifier with sensitivity = specificity = 0.99, andunequal class probabilities: PN = TN+FP = 0.9, PP =


TP+FN = 0.1 ⇒ 92% probability of a positive result beingcorrect (PPV=0.92), 0.1% probability of a negative resultbeing wrong (NPV=0.999)

• For sensitivity = specificity = 0.9 (and same PN, PP asabove), ⇒ 50% probability of a positive result being correct,but still only 1.2% chance of a negative result being wrong(PPV=0.5, NPV=0.988)

Slide 22: Fluoroscopy

• Provides real-time images

• X-ray source → patient → fluores-cent → recorder

• X-ray image intensifier (XRII)

– Cesium iodide phosphorus deposited directly on XRIIphotocathode

– Output image approximately 105 times brighter thaninput image

∗ flux gain (amplification of photon number) ≈ 100

∗ minification gain (from large input onto small out-put screen) ≈ 100

– quantum noise (small number of photons) limiting im-age quality

• Flat-panel detectors

– increased sensitivity to X-rays, reducing patient radia-tion dose

– Improved temporal resolution, reducing motion blurring

– Improved contrast ratio over image intensifiers

– Spatial resolution is approximately equal

Slide 23: C-arm

• A portable fluoroscopy machinethat can move around the surgerytable and make digital images forthe surgeon

• A limited number of projectionsis often used to reconstruct a 2-D“slice” through the 3-D volume

• High density objects (e.g. a nee-dle) or density gradients may cre-ate disturbing fan-shaped artifacts


Slide 24: Use of passive contrast agents

• Enhanced images may be made using a substance which isopaque to X-rays

• This is normally as part of a double contrast technique, usingpositive and negative contrast

• Positive radiographic contrast agents:

– Iodine (Z = 53) can be injected into bloodstream

– Barium (Z = 56)

• Negative radiographic contrast agents:

– air and carbon dioxide (CO2)

– CO2 is easily absorbed and causes less spasm

– CO2 can be injected into the blood, air can not!

Slide 25: Angiography

Used to visualize the inside of blood vessels and organs.Ancient

greek: angeion = “container”, grapho = “I write”

• Blood has the same X-ray density as surrounding tissue

• An iodine-based contrast is injected into the bloodstream andimaged as it travels around

• Angiography used to find:

– Aneurysms (abnormal blood-filled swellings of an arteryor vein, resulting from a localized weakness in the wallof the vessel)

– Leaks

– Thromboses (blood-clots that form and cause obstruc-tion of the blood vessel), etc

• The X-ray images may be:

– Still images, displayed on a fluoroscope or film

– Video sequences (25–30 frames per second)

• Retinal angiography: commonly performed to identify ves-sel narrowing in patients with e.g. diabetic retinopathy andmacular degeneration [retina = light sensitive tissue liningthe eye inner surface, macula = yellow spot near retina cen-ter responsible for central vision]

Slide 26: Digital subtraction angiography

• Angiography images: made while injectingcontrast medium into the bloodstream

– Image includes all overlying struc-tures besides the blood vessels

– Useful for determining anatomicalposition of blood vessels

• To remove distracting structures, a maskimage of the same area is acquired beforethe contrast is administered


• An image intensifier (fluoroscopy) is used, producing imagesat a 1 – 6 frames per second rate, subtracting all subsequentimages from the original “mask” image in real time

• Hence the term “digital subtraction angiography” (DSA).

• DSA is being used less and less, being taken over by CTangiography, which is less invasive and less stressful

Slide 27: Reconstruction from projections

• A 3-D object distribution can be mapped as a series of 2-Dprojections

• With a sufficient number of projections the mapping processcan be inverted and the 3-D distribution reconstructed fromthe projections

• If the projection axes lie in a plane, the reconstruction maybe carried out one slice at a time

• Then the inversion is simpler and may be done using a Fouriermethod

• To avoid a large matrix, a Maximum Likelihood (ML) methodis used and the solution is found by iteration

• Mathematical foundation presented by Johann Radon in 1917

Slide 28: Computed Tomography (CT) history

• CT as an imaging technique was described by Cormack in1963

• The first clinical implementation made by Hounsfield in 1972

• 1971 prototype made 160 parallel readings in 180 angles, witheach scan taking 5 minutes

• Image reconstruction from these scans took 2.5 hours

• Hounsfield and Cormac shared the 1979 Nobel Prize

Slide 29: Reconstruction radiography

• CT uses X-rays

• Instead of a 3-D cone beam, theyare collimated to travel in a 2-D“fan–beam”

• A 2-D projection of a cross section of the body is detectedby a large number of detectors

• Repeated for many orientations as the X-ray tube and thedetectors rotate around patient

• An image of the cross-section is then computed from theprojections


Slide 30: Tomographic reconstruction

• Data series collected: integrals at po-sition r, across a projection at angleθ

• Repetition for various angles

• Total attenuation given by line inte-gral:

pθ(r, θ) =

∫

A→B

µ(x(s), y(s))ds,

where all points on the line A → B satisfy r = x cos θ + y sin θ, so

pθ(r, θ) =

∫

∞

−∞

µ(x, y)δ(x cos θ + y sin θ − r)dxdy.

• This is the Radon transform of the 2-D object

• Projection-slice theorem:

– Infinite number of projections ⇒ perfect object recon-struction

– Inverse Radon transform ⇒ estimate of object functionµ(x, y)

– Unstable with respect to noisy data.

– Stabilized and discretized version: filtered back-projectionalgorithm

• In-depth theory: http://www.slaney.org/pct/pct-toc.html,ch. 3

Slide 31: 3 different CT modalities

• A single-slit CT is the simplest:

– Gives a single axial-plane image

– In axial “step and shoot”, the table is moved betweeneach slice

• In helical CT, the X-ray tube and the detectors rotate, whilethe patient is moved along an axis through the center ofrotation:

– Rapidly acquires 3-D data

– Slightly lower z-axis resolution than “step and shoot”

– May tilt detector ±30◦ relative to the axis of rotation

• In multislice CT, several rows of detectors gather a cone ofX-ray data, giving a 2-D projection of the 3-D patient

– With 1-3 revolutions per second, near real-time 3-Dimaging is possible, with translation speeds up to 20cm per second


Slide 32: CT advantages

• CT eliminates superimposition of structures outside ROI

• Small differences in physical density can be distinguished

• CT data can be viewed as images:

– in an axial plane

– in a coronal plane

– in a sagittal planes

– (multiplanar reformatted imaging)

• CT angiography avoids invasive insertion of an arterial catheterand guidewire

• CT colonography is as useful as a barium enema x-ray for de-tection of tumors, but may use lower radiation dose [En-ema: procedure of introducing liquids into the rectum andcolon via the anus, barium: used as a contrast agent]

Slide 33: CT disadvantages

• CT is a moderate to high radiation diagnostic technique:

– Improved radiation efficiency ⇒ lower doses

– Higher-resolution imaging ⇒ higher doses

– More complex scan techniques ⇒ higher doses

• Increased availability + increasing number of conditions ⇒large rise in popularity

• CT constitutes 7% of all radiologic examinations (UK)

• Contributed 47% of total collective medical X-ray dose

• Overall rise in total amount of medical radiation used, despitereductions in other areas

Slide 34: Contrast agent disadvantages

A certain level of risk associated with contrast agents

• Some patients may experience severe allergic reactions

• Contrast agent may also induce kidney damage (nephropa-thy):

– If normal kidney function, contrast nephropathy risknegligible

– Risk is increased with patients who have:

∗ preexisting renal insufficiency (= kidney failure)

∗ preexisting diabetes

∗ reduced intravascular volume.

– For moderate kidney failure, use MRI instead of CT

– Dialysis patients do not require special precautions:

∗ little function remaining

∗ dialysis will remove contrast agent.


Slide 35: 3-D surface / volume rendering

• Surface rendering:

– Threshold value chosen by operator (e.g. correspondingto bone), or a threshold level set using edge detectionalgorithms

– From this, a 3-dimensional model can be displayed

– different thresholds and colors may represent: bone /muscle / cartilage [Norwegian: brusk ]

– interior structure of each element is not visible in thismode

– will only display surface closest to viewer

• Volume rendering:

– transparency and colors are utilized

∗ bones could be displayed as semi-transparent

∗ one part of the image does not conceal another

Slide 36: Rendering examples

• Slices of a cranial CTscan (extreme right).

• Blood vessels are brightdue to injection of con-trast agent.

• Surface rendering showshigh density bones.

• Segmentation removesthe bone, and previouslyconcealed vessels cannow be demonstrated.

Slide 38: Nuclear medical imaging

• We need to generate contrast by local activity

• We inject a radioisotope carried by a molecule which is ab-sorbed differentially according to local metabolic rate

• Most of the radiation from the decay should escape the bodywithout attenuation or scatter

• Half-life of decay should match duration of procedure

• SPECT involves high patient dose, poor spatial resolution,and exceptional contrast

• PET solves most of SPECT’s shortcomings, but procedure ismore complex and expensive


Slide 39: Single photon emission computed tomography (SPECT)

• Mostly used for study of blood-flow, by injection of a radio-pharmaceutical into the bloodstream. May otherwise ingestor inhale the radiopharmaceutical

• Image obtained by gamma camera is a 2-D view of 3-D dis-tribution of a radionuclide

• SPECT imaging is performed by using a gamma camera toacquire multiple 2-D image

• Tomographic recon-struction yields 3-Ddataset

• From this dataset wemay show thin slicesalong any chosen axissimilar to MRI, CT,PET

Slide 40: Choice of radioactive isotope

• Radiation from decay should escape body:

– eliminates alpha radiation

• Minimal scattering to make sharp images:

– eliminates beta radiation

• Energy deposition should be minimal:

– eliminates gamma emission below 70 keV

• Half-life to match duration of procedure

• Short half-life to minimize radiation dose ⇒ radioactiveisotope w/ gamma half-life ≈ 103 s, energy ≈ 105 eV ⇒99Tc (Technetium), produced by β decay of99Mo (Molybdenum), produced in neutron-induced fission of 235U, or produced by neutronabsorption by 98Mo

Slide 41: SPECT cameras

• SPECT images are collected by pinhole gamma camera ro-tating around patient

• Projections are acquired at defined points during the rota-tion, typically every 3-6 degrees

• A full 360◦ rotation gives optimal reconstruction:

– 15 – 20 seconds / projection

– Total scan time: 15–20 minutes

• Multi-headed gamma cameras ⇒ faster acquisition:

– Dual-head camera give 2 projections simultaneously

– Triple-head cameras with 120 degree spacing are alsoused


Slide 42: SPECT reconstruction

• Low-resolution images (64×64 or 128×128 pixels)

• Pixel sizes: 3–6 mm

• Reconstructed images compared to planar images:

– lower resolution, increased noise, reconstruction arti-facts

• Uneven distribution of nuclides may also cause artifacts

• Attenuation ⇒ underestimation of activity with depth

– Modern SPECT equipment integrated with X-ray CTscanner

∗ CT images are attenuation map of tissues

∗ Incorporated into the SPECT reconstruction to cor-rect for attenuation

– Co-registered CT images provide anatomical informa-tion

Slide 43: Positron emission tomography (PET)

1. A sugar (fluorodeoxyglucose,FDG) containing radioactive 18Fis injected

2. During decay, isotope emitspositron

3. Positron annihilates with an elec-tron⇒ pair of γ photons mov-

ing in opposite direc-tions

• Gamma photons are detected by scanning device

• Only simultaneous pair of photons used for image reconstruc-tion

Slide 44: PET examples

Maximally intensity projection of typical full-body 18F: (GIFanimation file PET-MIPS-anim.gif) PET scan of the human

brain:


Slide 45: Orthogonal PET slices

Slide 46: Shift in PET application areas

Main PET areas in the 1990s

Main PET areas today:

• Cardiology: disorders of the heart and blood vessels

• Oncology: tumors (cancer), neurology: the nervous system

Slide 47: Oncology: lung case

• Traditionally: lung masses evaluated with X-rays, CT, and,more recently, MR. To determine malignancy: biopsy havebeen performed

• Now: PET can determine malignancy

55F, diagnosed w/ stage III “lung non-small cell carcinoma” (NSSC)


PET findings: Increased uptake of FDG in several lung nodules ⇒recurrent tumour indication. Also found: abnormal uptake in thecervical lymph nodes (neck lymph nodes)

Slide 48: PET benefits

• Principal benefit: sensitivity to imaging metabolic activity

• Spatial information:

– better than SPECT, worse than CT or MRI

• PET images may be fusedwith X-ray CT

– on the patient

– at the same time

– in the same machine

⇒ functional + anatomicalimage

Slide 49: PET limitations

• PET scanning uses short half-life isotopes: 11C (∼20 min),13N (∼10 min), 15O (∼2 min), and 18F (∼110 min)

• Timing and logistical limitations restrict clinical PET to 18F

• Frequent recalibration of remaining dose of 18F needed

• Ethical limitations to injecting radioactive material:

– short-lived radionuclides minimize radiation dose

– in cancer therapy, risk from lack of knowledge therapyresponse may be much greater than the risk from dueto PET radiation

• Isotopes must be produced in a cyclotron ⇒ high costs

Slide 51: Magnetic resonance imaging (MRI)

• Non-invasive method, based on nuclear magnetic resonance

• First utilized in physical and chemical spectroscopy

• 1971: Different relaxation times of tissues and tumors

• 1971: MRI first demonstrated on test tube samples

• 1973: first image published

• 1977: first image of human published


Slide 52: MRI history

• 1952: Bloch and Purcell: Nobel Prize for discovering Nuclearmagnetic resonance (NMR)

• 1975: Ernst: MRI using phase and frequency encoding, andFourier Transform

• 1977: Mansfield: developed the echo-planar imaging (EPI)technique

• 1987: Dumoulin: magnetic resonance angiography (MRA)

• 1991: Ernst: Nobel Prize in chemistry for pulsed FourierNMR and MRI

• 1992: functional MRI (fMRI) developed [imaging of changesin blood flow in the brain]

• 2003: Lauterbur and Mansfield: Nobel Prize for discoveriesin MRI

Slide 53: MRI basics

• Nuclei of hydrogen atoms (protons) align either parallel or

antiparallel a strong static external magnetic field ~B0

• Electromagnetic RF wave at resonance frequency υ transmitin a plane perpendicular to the magnetic field.

υ = γ|B0|, (1)

γ: gyromagnetic ratio. For hydrogen: γ = 42.58 MHz / T

• Puts nuclei in non-aligned high-energy state

• As protons return to alignment, they precess

• Precession generates new RF signal, picked up by antenna

Slide 54: Image formation

• To excite only protons in selected body parts: gradientsadded ⇒ spatial variation of ~B0(~r) ⇒ only selected partsof object resonate at the transmit RF υ

• Tomographic methods to generate 2-D images (back-projectionetc.)

• Stronger gradients permit faster imaging or higher resolution

• Faster switching of gradients permits faster scanning. Lim-ited by safety concerns over nerve stimulation

• Typical resolution: ∼1 mm3


Slide 55: The T1 process

• At equilibrium:

– Magnetization vector ~M of the pro-tons lies along the ~B0: equilibriummagnetization ~M0

– Mz: longitudinal magnetization

– No transverse magnetization Mx orMy.

• Time constant T1: describes how Mz re-turns to equilibrium: spin-lattice relax-ation time

Mz = M0

(

1 − e−t/T1

)

• T1: time to reduce difference between the longitudinal mag-netization Mz and its equilibrium value M0 by a factor ofe

Slide 56: The T2 process

• T2: describes return to equilibrium ofthe transverse magnetization, Mxy: spin-spin relaxation time

Mxy = Mxy0e−t/T2 (2)

• T2, T1 processes occur simultaneously

• Always: T2 ≤ T1

– Mxy → 0 as t → ∞,

– Mz → M0 as t → ∞

Slide 57: MRI contrast agents

• T1 / T2 images don’t always show anatomy/pathology. Maybe enhanced by injection of contrast agents

• Most common: paramagnetic contrast agents

– Appear extremely bright on T1-weighted images

– High sensitivity for detection of vascular tissues (e.g.tumors)

– Permits observation of brain perfusion (e.g. in stroke).

• Super-paramagnetic contrast agents (e.g. iron oxide nanopar-ticles):


– Appear dark dark on T2-weighted images

– Used for liver imaging: normal liver tissue holds backagent. Scars and tumors lets it in.

• Diamagnetic contrast agents:

– Barium sulfate for use in gastrointestinal tract

Slide 58: MRI modes

• Standard MRI: Including several different pulse sequences(time-series of different RF pulses to manipulate M)

• Echo-planar imaging (EPI): Each RF excitation excitation:followed by a train of gradients with different spatial encoding⇒ rapid data collection ⇒ less motion artifacts

• MR spectroscopy: Images other nuclei besides H: e.g. P(phosphorus), Na (sodium), F (fluorine)

• Functional MRI (fMRI): Measures change in blood-flow inthe brain. Hemoglobin: diamagnetic when oxygenated, para-magnetic when deoxygenated ⇒ magnetic resonance signalof blood is different depending on oxygenation level

Slide 59: MR angiography (MRA)

• MRA used to image blood vessels to evaluate for:

– Stenosis: abnormal narrowing

– Aneurysms: localized, balloon-like blood-filled bulge (mostcommonly in arteries)

• Techniques used are e.g.:

– Injection of paramagnetic con-trast agents

– “Flow-related enhancement”:Stationary tissue: has been inimaging plane for long time⇒ not fully “relaxed” beforenext RF excitation ⇒ respondsdifferently than “fresh” bloodthat just entered the imaging

plane

Slide 60: Magnetic resonance spectroscopy (MRS)

• MRI shows the location of a tumour

• MRS indicates how how aggressive (malignant) it is


• MRS can be tuned to different chem-ical nuclei, e.g. H (hydrogen), P(phosphorus), Na (sodium), F (fluo-rine)

• MRS used to investigate

– Cancer (brain / breast /prostate)

– Epilepsy, Parkinson’s, andHuntington’s (a neurodegenera-tive genetic disorder)

• MRS example: [University of Hull, Centre for

Magnetic Resonance Investigations]

5 mm thick axial MRI brain slice (tu-mor at bottom right)Red box: region of interest for MRS

• Proton MRS spectrum from markedregion of interest:Red peaks correspond to alanine, gen-erally only seen in meningiomas

Slide 61: Functional MRI (fMRI)

• Changes in brain activity linked to:Changes in blood flow and blood oxy-genation

• Active nerve cells consume oxygencarried by hemoglobin

• Hemoglobin is:

– Diamagnetic when oxygenated

– Paramagnetic when deoxy-genated

⇒ MR signal of blood depends on level of oxygenation ⇒ “Blood-oxygen-level dependent (BOLD) contrast”


Slide 62: Intraventional MRI

• MRI scanner used to si-multaneously guide minimally-invasive procedure:

– Strong magnetic radiofre-quency field present

– Quasi-static fields gener-ated⇒ Non-magnetic environ-ment, instruments, andtools required

• Open magnet gives surgeonbetter access to patient (seeimage)

⇒ Often implies lower field magnets (∼0.2 T) ⇒ Decreased sensi-tivity

Slide 63: Current density imaging (CDI)

Used for mapping electric current pathways through tissue

1. External current is applied during MRI session

2. Electrical currents generate local magnetic fields

3. Such magnetic fields affect the phase of the magnetic dipolesduring an imaging sequence

4. CDI uses phase information from MRI to reconstruct currentdensities within the object

[Data courtesy of NYU Medical Center]

Slide 64: MRI vs. CT

• CT: differentiates high Z tissue (bone, calcifications) fromcarbon based flesh

• MRI: best suited for non-calcified tissue

• CT: may be enhanced by contrast agents containing highatomic-number atoms (e.g. iodine, barium)

• MRI: Contrast agents have paramagnetic / diamagnetic prop-erties


• CT: utilizes only X-ray attenuation to generate image con-trast

• MRI: a variety of properties that may generate image con-trast

• CT: usually more available, faster, much less expensive

• MRI: generally superior for tumor detection and identifica-tion

• MRI: best if patient is to undergo examination several times

• CT: if repeated, may expose the patient to excessive ionizingradiation

Documents

Medical Imaging Methods, in Brief - Universitetet i oslo€¦ · ter responsible for central vision] Slide 26: Digital subtraction angiography • Angiographyimages: made while injecting